Notes on the Scientific Method: Observation, Reliability, and Pseudoscience
Scientific Method: Living Things, Evidence, and Public Trust
- Life science = study of living organisms; uses the scientific method to study living systems.
- The scientific method is a set of rules developed over decades that has helped answer questions about how the world works.
- Scientists report their findings; the public evaluates and may call for changes based on observations.
- The scientific process is not linear; relationships do not automatically imply causation. Correlation ≠ causation.
Non-linearity, Causation, and Interpreting Results
- Relationships in science can be linear or non-linear; complexity makes causal inference difficult.
- When two researchers study the same topic, they can obtain different results due to various factors (examples and biases discussed below).
- The wine example: two subjects in a study on wine can yield different outcomes; highlights variability in human results and the influence of context.
- Biases and design differences can drive divergent results:
- Sampling bias: one group more biased in participant selection.
- Health differences in participants (e.g., age ranges like 75–80) and volunteer bias.
- Control group differences and whether the control group truly matches the experimental group.
- Sponsorship or financial interests can influence results.
- Family history (genetics) can influence results (e.g., hereditary heart disease).
Reliability, Replication, and Evidence Synthesis
- If two studies have opposite conclusions, multiple factors could be at play.
- Science seeks reliability through replication and accumulation of evidence.
- If you have many studies showing a similar result (e.g., humans are warming the planet over the last ~${$10^{2}}$} years), a single opposing study is weighed against the larger body of evidence.
- The repeatability of results strengthens scientific confidence; more studies confirming a result increases trust in that finding.
- Example from the speaker’s field: conflicting studies on how bird size affects pitch, illustrating that not all questions have a single, settled truth yet.
The Iterative, Competitive Process of Science
- Science is iterative and competitive: many studies ask the same questions, repeat experiments, and accumulate evidence.
- Over time, this process can converge toward a scientific truth for certain questions.
- If there is incomplete understanding or conflicting results, more studies are conducted to clarify.
Components of the Process of Science
- Observation: careful watching to gain information.
- Examples: hotter summers in August vs December; earlier sunsets in July vs March.
- Hypothesis: a testable explanation for observed phenomena.
- Hypotheses can arise from logical reasoning, experience, chance, intuition, or established theory.
- Examples:
- Hypothesis from experience: being cold and wet after swimming causes illness.
- Literature examples: cuttlefish skin showing millions of colors without a power source, suggesting inspiration for non-powered display tech (hypothetical).
- Hypotheses should be testable and supported by data; they should explain phenomena in a valid, measurable way.
- The hypothesis process can blend creative and logical thinking.
- Literature review: examine existing reports and findings before forming new tests.
- Testability and falsifiability:
- Testable hypothesis: can be supported or refuted by data.
- Non-testable (non-falsifiable) hypothesis: cannot be measured or disproven (e.g., psychic energy explanations).
- Distinguishing between testable and non-testable is crucial; science can lean on philosophy for non-testable questions.
- Pseudoscience vs science:
- Pseudoscience features: vague or babble-like language; lack of a coherent mechanism; extraordinary claims with insufficient evidence.
- Science prizes evidence: receipts, data, and reproducible results.
- Pseudoscience often lacks peer-reviewed validation and relies on non-testable claims.
- Falsifiability and calcifiability:
- A falsifiable (calcifiable) hypothesis can be disproven by evidence.
- A non-falsifiable hypothesis cannot be proved wrong and is not scientifically testable.
- Peer review:
- Scientific claims are evaluated by peers in the same field before publication.
- Peers check for flaws, repeat experiments if needed, and provide validation or critique.
- Conspiracy claims vs evidence:
- Scientific practice requires testability and verifiability (receipts/evidence).
- Conspiratorial claims are not considered scientific without testable evidence and reproducibility.
Practical Implications for Evaluating Scientific Claims
- When faced with conflicting results, seek additional studies and replication rather than accepting a single study.
- Consider potential biases: sampling bias, control group differences, volunteer bias, sponsorship bias, and social biases.
- Recognize that science is not about certainty in every case; it is about the weight of evidence built through repeated testing and convergence of results.
- Acknowledge the role of philosophy in questions that are not readily testable by current methods (e.g., psychic energy).
- Remember the maxim frequently invoked in science: show me the receipts. Evidence, reproducibility, and peer review are essential.
Real-World Takeaways and Examples Mentioned
- The wine study controversy illustrates how different subject pools can yield different results.
- The smoking example highlights long-term effects and timescales (roughly a decade) used to assess health outcomes.
- The broader point that data, statistics, and the mean (average) help summarize effects across groups:
- Mean is often used to describe central tendency: ar{x} = \frac{1}{n}\sum{i=1}^{n} xi
- The discussion about larger birds and pitch demonstrates that not all scientific questions have a settled answer; more studies can resolve some questions while others remain open.
- Cuttlefish color patterns as an example of how literature can inspire real-world technology and hypotheses (even if the example is hypothetical in this context).
- The role of replication: multiple studies addressing the same question increase confidence in findings; a single contradictory study is not sufficient to overturn a broad consensus.
- The overall message: science is an iterative, evidence-based, and public enterprise that advances by building reliability, addressing biases, and differentiating testable from non-testable questions.